HIV testing · PLoS Med · 2014
Home HIV testing won on a main goal that the trial’s public list never included.
Each goal’s place on the trial’s public list and in the paper
The paper says
A trial in rural Lesotho offered HIV tests at home or a travelling clinic. Its main goal: what share of those offered a test took one. Home won, 92.5% to 86.7%. The paper says it dropped one publicly listed goal and “kept” this one.
Its own data show
No version of that list, from October 2011 to February 2012, names getting tested. Results on its first two goals, from the paper’s own table: finding undiagnosed HIV (clinic 6.2% of those tested, 75 of 1,207; home 3.6%, 39 of 1,083) and getting them into care within a month (clinic 25.3%, home 25.6%).
Trials post their main goals, dated, on a public website that keeps every edit, so readers can check a winning goal wasn’t chosen after seeing the results. This trial’s own planning document does put getting tested first, but the paper never says the public list left it out. On the list’s first two goals, home testing was no better.